html_url,issue_url,id,node_id,user,created_at,updated_at,author_association,body,reactions,performed_via_github_app,issue
https://github.com/pydata/xarray/issues/4927#issuecomment-782783402,https://api.github.com/repos/pydata/xarray/issues/4927,782783402,MDEyOklzc3VlQ29tbWVudDc4Mjc4MzQwMg==,2448579,2021-02-21T02:22:02Z,2021-02-21T02:22:02Z,MEMBER,"Thanks @spencerkclark I missed that those conventions specified that the `units` attribute should include the datetime string.

@kmuehlbauer you could
``` python
ds = xr.open_dataset(..., decode_times=False)
ds = rewrite_time_reference_units(ds)
ds = xr.decode_cf(ds)
```","{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,811933067
https://github.com/pydata/xarray/issues/4927#issuecomment-782111696,https://api.github.com/repos/pydata/xarray/issues/4927,782111696,MDEyOklzc3VlQ29tbWVudDc4MjExMTY5Ng==,2448579,2021-02-19T14:32:28Z,2021-02-19T14:32:28Z,MEMBER,"https://github.com/pydata/xarray/blob/070d8158aab31c6a53efb36f9ea4a57dc6671e43/xarray/conventions.py#L498-L500

Looks like we treat time bounds specially. You could add a helper function to this if condition that rewrites `""seconds since time_reference""` to `""seconds since 1970-01-01T00:00:00Z""` and the rest of the decoding machinery should just work. 

This approach might break round-tripping that file though. Perhaps saving `encoding[""units""] = ""seconds since time_reference""` fixes that?","{""total_count"": 1, ""+1"": 1, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,811933067